Optimization and Simulation of Platform Supply Pickup and Delivery - Case from the Brazilian Petroleum Industry
MetadataVis full innførsel
The oil industry has a vital role in energy provision and the economic aspects of its operations, which is associated with high values and risks. Continuous oil production from fields is essential, and it is important that this is supported by robust logistics solutions. Offshore production facilities require supplies that are transported from an onshore port to the facility by platform supply vessels. Ship transportation is the most costly part of the upstream logistics sector, making good planning even more critical. This study examines the order cycle for Petrobras in the Campos Basin in Brazil. Models are developed simulating the situations where supply orders are generated from a random distribution and where different policies concerning the vessel voyages are applied. Periodic problems are solved to determine the order service and corresponding sailing routes for each given ship journey. The objective functions minimises costs for orders that are not served at departures from where they are requested, in addition to the distances travelled when this is proposed as a possibility. Inconvenience costs were set proportional to the demand quantities for the specific orders, with more weight put on delivery service rather than pickup. The model is flexible and simple with fixed routes and schedules in use, imitating the present situation. Challenging the voyages however, complicates it. Petrobras experience problems in their supply chain originating in onshore logistics. 25 % of the orders scheduled for given departures arrive at the port too late to be transported, and 50 % of orders are requested as emergencies whilst the percentage in fact should be 10. No-shows lead to low utilisation of the vessels and frequent express leasing, which if further impacted by the amount of emergencies in need of hurried services with express. Overbooking is a strategy that potentially can correct parts of the problem. This was proven when planning for 20-25 % excess capacity than the actual, where the expected delays were reduced by one third. If Petrobras were able to reduce the proportion of no-shows to 15 % and emergency requests back to its normal state of 10 %, the analyses show reductions of nearly 40 % and 80 % in order delay and express demand, respectively. In addition, if they were able to make their Logistics compatible with dynamic route planning as well, the potential of saving up to 70-75 % in order delay is evident, where there no longer exists any reasons to use express. This should act as an incentive to make an effort to improve the logistics system. Nevertheless, there is a trade-off with the use of different policies applied due to the uncertain nature of the problem. A detailed evaluation is left to the decision makers.